In hydrodynamics there are two basic approaches. The first is the Eulerian specification where the coordinate system is fixed. In that case, the partial time derivative and volume integral operators can be interchanged: $$ \int_V\frac{\partial f}{\partial t}dV=\frac{\partial}{\partial t}\int_VfdV, $$ the volume being fixed in space. The second specification is the Lagrangian one where you follow one fluid "particle" and the relevant time derivative is the Lagrangian/material derivative: $$ \frac{\partial f}{\partial t}+\mathbf{u}\cdot\nabla f\equiv\frac{Df}{Dt}, $$ where $\mathbf{u}$ is the flow field. In that case, the Lagrangian derivative can be interchanged with the volume integral, but only if the volume is that of the fluid "particle": $$ \int_{V_u}\frac{Df}{Dt}dV_u=\frac{D}{Dt}\int_{V_u}fdV_u. $$
Now I've seen in a paper the author using this latter interchange of operators, but for a fixed volume comprising of the entire system (actually the volume going up to infinity). Is this valid? How do you make that step? I couldn't find any info on this. Maybe you just sum all the individual volume integrals and do the interchange with all of them? And then what happens to the Lagrangian derivative?